Fields Medalists Accuse AI Labs of Threatening Mathematics as OpenAI Feud Escalates

The open letter, published this week, follows a series of incidents that have frayed relations between mathematicians and AI companies. This week, NYU professor Tristan Buckmaster accused OpenAI of pressuring him not to credit a collaborator who works for Anthropic for solving an important math problem. Buckmaster also raised concerns that OpenAI may have used his team's work with their Codex tool to produce its own ground-breaking proof after a marathon weekend of inference.

On Thursday, OpenAI withdrew its sponsorship of a math event at CalTech after researchers at the university criticised the company's actions. OpenAI's own proof, announced recently, remains unverified by the broader mathematical community.

The signatories argue that while AI's ability to solve outstanding mathematical challenges could be a boon, such benefits will only materialise if solutions can be understood and communicated. "Often these solutions are announced in a rush, leaving no time for a proper writeup, the isolation of new methods and ideas, and citing relevant previous work of others," they wrote. "As in all creative professions, this raises severe attribution and plagiarism questions."

The letter also warns that without the willing engagement of mathematicians to integrate AI-conceived ideas into the canon, "the crucial human transmission chain between mathematicians would be lost." Mathematicians are growing paranoid that their use of AI tools may be fed back into new models, incentivising secrecy as frontier labs can now spend tens of millions of dollars using LLMs to beat original researchers to proofs.

The open letter follows the Leiden Declaration, released by a working group of mathematicians in June, which offered recommendations for mathematicians, institutions, and policymakers grappling with the changes LLMs bring to mathematical work. The signatories cautioned that the issues facing mathematics foreshadow similar disruptions in other scientific and creative professions, and ultimately for all of humanity.

Why This Matters

The feud underscores a fundamental tension between the speed of AI-driven discovery and the deliberate, peer-reviewed norms of academic mathematics.
If mathematicians retreat into secrecy to protect their work, the open research culture that has driven progress for decades could collapse.
The dispute sets a precedent for other fields—if AI labs can outrun human researchers and claim credit, similar conflicts are likely in physics, biology, and the humanities.

Background

Mathematics has long operated on a model of open collaboration, where credit and citation are the currency of career advancement. AI labs like OpenAI and Anthropic have increasingly used large language models to tackle long-standing problems, often producing results faster than traditional researchers can verify. The Leiden Declaration from June 2026 was an early attempt to address these challenges, but recent events—including Buckmaster's accusations and the CalTech sponsorship withdrawal—have brought the conflict to a head. The Fields Medal, awarded to mathematicians under 40, is the highest honour in the field, making the signatories' collective weight significant.

Key Perspectives

Academic Mathematicians: They seek to preserve the integrity of their discipline, insisting that proofs must be verified, understood, and integrated into the broader mathematical canon. They fear being reduced to data providers for AI models. AI Labs (OpenAI, Anthropic, etc.): They view AI-assisted proofs as a legitimate—and powerful—way to advance human knowledge quickly. They argue that speed and discovery should not be penalised, and that attribution norms must evolve. Critics/Skeptics: Some observers worry that mathematicians are protecting professional turf rather than genuinely defending rigour, and that AI could democratise problem-solving beyond the elite. Others point out that without proper verification, AI proofs could contain errors that go undetected for years.

What to Watch

Whether OpenAI's disputed proof is independently verified or retracted by the mathematical community.
The response from other AI labs—whether they adopt the Leiden Declaration's recommendations or ignore them.
Potential fallout for AI companies' partnerships with academic institutions, and whether similar open letters emerge from other scientific fields.

ZOTPAPER